There's no such thing as "intrinsic linearity," or at least I've never gotten a satisfactory definition of that oft-tossed-about term. A tube just lays there if it's not in a circuit. The sort of circuit you use will determine linearity. For example, if you say, "common cathode, SE, no feedback," the triode circuit will generally be more linear as a voltage amp. Unless you want the amp to work as a transconductance (V-to-I) amp, in which case the pentode circuit will be more linear. When you say, "common anode," the pentode will be more linear, whether in SE or push-pull. So which device is more "intrinsically linear"?
If you consider the entire amp as a system, and go back to the common cathode connection, push-pull, the open-loop distortion of a pentode stage may be actually pretty comparable (in magnitude if not order) to a triode stage. For example, a 6L6GC in pentode push-pull can yield better than 2% distortion open-loop. But you don't run a pentode amp open-loop as a voltage amp, you use its gain to give you a decent amount of feedback, enough (if you're doing a good design) to drop the magnitude of the higher-order nonlinearities to below any conceivable threshold of detection with the intended load. That amount of feedback just isn't available for a triode and it will be difficult to avoid significant levels of distortion, as well as bandwidth reduction (triodes have Miller effects and can actually increase the distortion of the driver stage; pentodes, not so).
That all said, I use triodes in my amps, though not necessarily conventional ones. Everything is a trade-off, and there have been great amps made with any output configuration you'd care to name.
If you consider the entire amp as a system, and go back to the common cathode connection, push-pull, the open-loop distortion of a pentode stage may be actually pretty comparable (in magnitude if not order) to a triode stage. For example, a 6L6GC in pentode push-pull can yield better than 2% distortion open-loop. But you don't run a pentode amp open-loop as a voltage amp, you use its gain to give you a decent amount of feedback, enough (if you're doing a good design) to drop the magnitude of the higher-order nonlinearities to below any conceivable threshold of detection with the intended load. That amount of feedback just isn't available for a triode and it will be difficult to avoid significant levels of distortion, as well as bandwidth reduction (triodes have Miller effects and can actually increase the distortion of the driver stage; pentodes, not so).
That all said, I use triodes in my amps, though not necessarily conventional ones. Everything is a trade-off, and there have been great amps made with any output configuration you'd care to name.
Hi,
Sure.
Which is exactly why such statements as Tim's shouldn't be made without a specific qualifier for it's absolutely worthless without it.
Every single designer working with tubes across the world I know of works with the data given on datasheets, curves, loadlines etc.
When they say triode "X" is more "linear" than triode "Y" it is understood they mean as a voltage amplifier without any form of external FB applied.
Of course those nice triode curves that look sooo linear are absolutely useless if you deliberately decide to work with a (there you have it again) "non-linear" part of them.
But no, Mr. Einstein feels he's entitled to blanket statements with his fingers stuck firmly into both his ears since he's now turned eighteen.
Sorry but there's just nothing to defend when all conventions are thrown out and everyone's allowed "to think outside the box" when it best suits them without as much as a word of explanation.
Cheers,
Sure.
Which is exactly why such statements as Tim's shouldn't be made without a specific qualifier for it's absolutely worthless without it.
Every single designer working with tubes across the world I know of works with the data given on datasheets, curves, loadlines etc.
When they say triode "X" is more "linear" than triode "Y" it is understood they mean as a voltage amplifier without any form of external FB applied.
Of course those nice triode curves that look sooo linear are absolutely useless if you deliberately decide to work with a (there you have it again) "non-linear" part of them.
But no, Mr. Einstein feels he's entitled to blanket statements with his fingers stuck firmly into both his ears since he's now turned eighteen.
Sorry but there's just nothing to defend when all conventions are thrown out and everyone's allowed "to think outside the box" when it best suits them without as much as a word of explanation.
Cheers,
That may or may not be so, but there's nothing that privileges one sort of linearity versus another, i.e., there's nothing intrinsic involved, just a convention of jargon. And getting too hung up on convention can (without one meaning to) limit one's view of design options. You've seen me get cranky before when people use terms like "bias" and "idle current" interchangeably and this is sort of the same thing to me- it's an impediment to broad understanding for too many people.
Hi,
As far as I can tell it goes both ways and from what I've seen so far the people involved here adhere to the same conventional "wisdom" or lack thereof as I would.
If we'd care to return to page #1 of this thread it may become clear why I get cranky about the pointless statements as quoted from in the above postings.
Do we want to convince everyone that pentodes are more linear and hence in most peoples minds the devices "par excellence" for the topic on hand here or shall we re-explain the meaning of words such as curves, linear, transfer, feedback or what have we?
Or am I just preaching to the choir??
Cheers,
As far as I can tell it goes both ways and from what I've seen so far the people involved here adhere to the same conventional "wisdom" or lack thereof as I would.
If we'd care to return to page #1 of this thread it may become clear why I get cranky about the pointless statements as quoted from in the above postings.
Do we want to convince everyone that pentodes are more linear and hence in most peoples minds the devices "par excellence" for the topic on hand here or shall we re-explain the meaning of words such as curves, linear, transfer, feedback or what have we?
Or am I just preaching to the choir??
Cheers,
O.K. there is no "formal" definition of intrinsic linearity, but most people would agree that a "less parabolic-shaped" dynamic characteristic curve would indicate "intrinsically more linear". If Dr. Williams is right, then why triodes?
John
P.S. I had a Sherwood receiver in the 80's that had 0.00001% distortion, which proves that cheap transistors are the most linear devices.
Because Dr. Williams is only right within a limited domain, just like the people who say things like "triodes are more linear than pentodes." Linearity is a property of a circuit, not of a device. The proper device to optimize the linearity of a circuit depends on the circuit. It just ain't that complicated. That's why there's more than one kind of tube.
And you've fallen into the trap of unstated assumptions as soon as you use the term "dynamic characteristic curve." Which one? Ip versus Eg at constant Ep? Ep versus Eg at constant Ip? Ep versus Eg2 at constant Eg1? And so on and so on. You'll get a different answer depending on which transfer characteristic you're talking about; which curve is relevant depends on what the circuit requires from the tube and we're back to what I said in the above paragraph.
Intrinsic linearity IS the linearity of the device.
Ip vs. Eg at constant Ep. Although it really doesn't matter, because triodes are intrinsically more linear amplifying devices than pentodes.
Your argument seems also to imply that triodes can be more efficient than pentodes, simply because you can operate a pentode inefficiently.
John
jlsem said:
Dude, look at the fricking nonlinear local feedback through the screen! It's right there under your nose. Or in a pure triode, through the electrostatic field. The screen stops that and allows it to be applied MORE LINEAR with local grid or global NFB, and more gain as well.
It's not a better distortion spectrum, but you can't hear any of it below 0.1% so no one cares.
Tim
In my first amp, I have 6Bm8 wired SE pentode. Then I realised my transformer can be wired ultra-linear mode, I gave it a try.
Now its SE-UL, and i was never happier before
Big difference in distortion, and sound balance. Female voices used to sound to thick to my taste, now its 'just right'
Now its SE-UL, and i was never happier before
Big difference in distortion, and sound balance. Female voices used to sound to thick to my taste, now its 'just right'
I think there will be no end for this. How does a circuit sound depends on the tube. I guess some tubes perform their best when employed in the right circuitary. There are no such thing that PP is better than SE. So i think the better question is given a specific tube, what is the best way to use it?
just my 3 cents opinion
just my 3 cents opinion
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.